Filtros Eletrônicos

ELECTRONIC FILTERS
Celso José Faria de Araújo, Dr.

CONCEPTS
✓ A “Ideal Electronic Filters” allow distortionless
transmission of a certain band of frequencies and suppress
all the remaining frequencies of the spectrum of the input
signal.
✓ The frequency spectrum is a representation of amplitude
versus frequencies of this signal.
1
Electronic Filters
Electrical Circuits - Celso José Faria de Araújo, M.Sc.
CLASSIFICATION
Electronic Filters
Classification Classification Classification

in terms of in terms of function in terms of
Technology Used accomplished Response-Function
Passive Low-Pass Butterworth

Filter Filter Filter
Active High-Pass Chebyshev

Filter Filter Filter
Digital Band-Pass Eliptic or

Filter Filter Cauer Filter
Band-Reject
Filter
2
Electronic Filters
FUNCTION ACCOMPLISHED
H ( j ) H ( j )
Ideal
Ideal K
K Real
K K
Real 2
2
c c 

(log scale) High-Pass Filter (log scale)
Low-Pass Filter
H ( j ) H ( j )
Ideal
Ideal
K K
K Real K
2 2 Real
c1 o c2  c1 o c2 

Band-Pass Filter (log scale) Band-Reject Filter (log scale)
3
Electronic Filters
ACTIVES FILTERS
❖Advantage:
✓ Suppress use of inductors;
✓ Easy Design of Complex Filters by cascade of simple stage;
✓ A considerable amplification of input signal is possible;
✓ Design Flexibility.
❖Disadvantages:
✓ Power supply is necessary;
✓ The frequency response is limited by active devices (Op-Amps,
Transistors) frequency response;
✓ It’s not often apply in medium and high power system.
Despite these disadvantages it’s widely used in several application, such
as: telecommunication and industrial instrumentation.
4
Electronic Filters
STANDARD REQUIREMENTS FOR
DESIGN OF APPROXIMATION
FUNCTION
LOW-PASS FILTER
Loss dB
Amin
Transition
Passband
Passband
Stopband
Amax
p s 
(log scale)
5
Electronic Filters
NON-FLAT STOPBAND
REQUIREMENTS FOR LOW-PASS
FILTER
Loss dB
Amin2
Amin1
Passband
Amax
p s1 s2 
(log scale)
6
Electronic Filters
FUNCTION
HIGH-PASS FILTER
Loss dB
Amin
Stopband
Transition Passband
Passband
Amax
s p 
(log scale)
7
Electronic Filters
NON-FLAT STOPBAND
REQUIREMENTS FOR HIGH-PASS
FILTER
Loss dB
Amin1
Amin2
Passband
Amin1
Amax
s2 s1 p 
(log scale)
8
Electronic Filters
FUNCTION
BAND-PASS FILTER
o 2 = 1 2
Loss dB
B =  2 − 1
o = 3 4
2
Amin
Transition Transition
Passband Passband
Stopband Passband Stopband
Amax
3 1 2 4 
(log scale)
• B is the passband width of BP filter

• o is the center (geometric mean) of the
passband of BP filter 9
Electronic Filters
NONSYMMETRICAL REQUIREMENTS
FOR BAND-PASS FILTER
o 2 = 1 2
Loss dB
B =  2 − 1
o = 3 
2 '
4
Amin2
Amin1
Passband
Amax
3 1 2 4’ 4 
(log scale)
• B is the passband width of BP filter

passband of BP filter 10
Electronic Filters
FUNCTION
BAND-REJECT FILTER
o 2 = 1 2 o 2 = 3 4 B =  2 − 1
Amin
Stopband
Transition Transition
Loss dB
Passband Passband
Passband Passband
Amax
1 3 4 2 
(log scale)
• B is the passband width of BR filter

stopband of BR filter
11
Electronic Filters
NONSYMMETRICAL REQUIREMENTS
FOR BAND-REJECT FILTER
o 2 = 1 2 o =  '  4 B =  2 − 1
2
3
Loss dB
Amax1
Amin
Amax2
1 3’ 3 4 2 
(log scale)
• B is the passband width of BR filter

stopband of BR filter
12
Electronic Filters
A SECOND-ORDER GAIN FUNCTION
FOR LOW-PASS FILTER
V  p2
GAIN = O =
VIN p
s2 + s +  p2
Qp
j
Loss dB
Slope = 40dB/decade
Gain
0dB
p 
(log scale)
13
Electronic Filters
FOR HIGH-PASS FILTER
V s2
GAIN = O =
VIN p
s2 + s +  p2
Qp
j
Loss dB
Slope = -40dB/decade Gain
0dB
p 
(log scale)
14
Electronic Filters
FOR BAND-PASS FILTER
p
s
VO Qp
GAIN = =
VIN p
s +
2
s +  p2
Qp
j
Loss dB
Slope = 40dB/decade Gain
Slope = -40dB/decade
0dB
p 
(log scale)
15
Electronic Filters
FOR BAND-REJECT FILTER
V s 2 +  p2
GAIN = O =
VIN p
s2 + s +  p2
Qp
j
Loss dB
Gain
p
0dB
p 
(log scale)
16
Electronic Filters
FOR LOW-PASS NOTCH FILTER
V s 2 +  z2
GAIN = O = ;  z   p
VIN p
s2 + s +  p2
Qp
j
jz
Loss dB
Gain
p
p z 
(log scale)
-jz
17
Electronic Filters
FOR HIGH-PASS NOTCH FILTER
V s 2 +  z2
GAIN = O = ;  z   p
VIN p
s2 + s +  p2
Qp
j
Loss dB
Gain
jz
p
-jz
z p 
(log scale)
18
Electronic Filters
GAIN EQUALIZERS
19
Electronic Filters
DELAY EQUALIZERS
vo (t ) = vin (t − To ) Vo ( s) = Vin ( s) e − sTo
Vo ( s )
H (s) = = e − sTo H ( j ) = e− j To
Vin ( s )
H ( j ) = 1  H ( j ) = −  To =  ( )
d
delay = (−  ( ))
d
20
Electronic Filters
DELAY EQUALIZERS
Second-order (all-pass)
p
s2 − s +  p2
VO Qp j
GAIN = = Gain
VIN p
s2 + s +  p2
Qp
p
     
 − p    p 
 Qp   Qp  
H ( j ) = 0 dB  H ( j ) = arc tg  2 2 
− arc tg  2 =  ( )
p − p −2 
   
   
   
d p  p2 +  2
delay = (−  ( ) ) = 2
d Qp 
2

( 2
p − 2
)
2
+ p
Q
 2

 p 
21
Electronic Filters
EQUALIZATION OF
CABLE DELAY
22
Electronic Filters
Filter pole-zero patterns
23
Electronic Filters
Bode Plot Approximation Technique
106
H ( s) =
35dB ( s + 100)(s 2 + 100s + 1002 )
Loss dB
3dB
100 400 
3 dB → 35 dB  35 dB
Loss (dB)
100 rad / s → 400 rad / s  2 oitava
order = 3
K
H (s) =
 
(s + 100) s 2 + 100 s + 1002 
 Qp 
K
H (s) s = 0 = 2
= 1  K = 106
100 100
Frequency (rad/s)
106
20 log = − 3  Qp = 1
 
(s + 100) s 2 + 100 s + 1002 
 Qp  s = j100
24
Electronic Filters
BUTTERWORTH APPOXIMATION
Requirements: p → Amax s → Amin
 100,1 Amin − 1 
log 0,1 Amax 
 10 − 1 
Order: n = 2
Correction Factor:  = 100,1Amáx − 1
 
log s 
  Roots of a third-order normalized filters
 p
Gain j
Roots of the normalized filters:
   2 k + n −1  
j  
Sk = e  2 n 
onde k = 1, 2, ...., 2n
n /3
1/H(S)
1 S+1 
-1 1
2 S2 + 1,414S + 1
3 (S2 + S + 1) (S+1)
4 (S2 + 0,76537S + 1) (S2 + 1,84776S + 1)
5 (S2 + 0,61803S + 1) (S2 + 1,61803S + 1) (S + 1)
  1n 
Denormalization: S = s  The poles are just the roots at left half plane
 p 
  25
Electronic Filters
Magnitude response for Butterworth filters of various order with  = 1. Note that as the order increases, the response approaches the
ideal brickwall type transmission.
26
Electronic Filters
Graphical construction for determining the poles of a Butterworth filter of order N. All the poles lie in the left half of the s-plane on a
( )
circle of radius 0 = p(1/)1/N, where  is the passband deviation parameter :  = 10 Amax /10 − 1
(a) the general case, (b) N = 2, (c) N = 3, (d) N = 4. 27
Electronic Filters
BUTTERWORTH APPOXIMATION
Example
Requirements: p = 100 rad/s Amax = 0.5 dB s = 400 rad/s Amin = 12 dB
 100,1x12 − 1 
log 0,1x 0.5 
 10 −1
order : n = 2
= 1.73 = 2
 400 
log 
 100 
Correction factor :  = 100,1x 0.5 − 1 = 0.35
1
Normalized H ( S ) =
S 2 + 1.414S + 1
  1n 
Denormalization : S = s  = 0.0059s
 p 
 
28727.4
Denormalized H ( s ) =
s + 239.6 s + 28727.4
2
28
Electronic Filters
CHEBYSHEV APPOXIMATION
 100,1 Amin − 1 
cosh  −1  Roots of a third-order normalized filters
 100,1 Amáx − 1 
Order : n =   where  =  s
cosh−1 ( s ) p
s
j 1 1
cosh sinh−1 
Gain n 
Roots of the normalized filters:
1 1
sinh sinh−1 
S k =  k + j k where k = 0,1,2,..., 2n − 1 n 
   1 + 2k   1 −1 1 
 k =  sen   senh senh 
 2  n  n 

   1 + 2k   1 −1 1 
 k =  cos   cosh senh 
 2  n    n    = 100,1A − 1
máx
s
Denormalization : S =
p The poles are just the roots at left half plane
29
Electronic Filters
Sketches of the transmission characteristics of a representative even- and odd-order Chebyshev filters.
30
Electronic Filters
(Order from Plot)
 
Requirements: p → Amax s → Amin =  s = s   p = 1
p p
Loss of LP Chebyshev approximation for Amax= 0.25dB Loss of LP Chebyshev approximation for Amax= 0.50dB Loss of LP Chebyshev approximation for Amax= 1dB
31
Electronic Filters
(Polynominal from Table)
n Numerator of 1/H(S) Amax= 0.25dB Denominator (k)
1 S + 4.10811 4.10811
2 S2 + 1.79668S + 2.11403 2.05405
3 (S2 + 0.76722S + 1.33863) (S+0.76722) 1.02702
4 (S2 + 0.42504S + 1.16195) (S2 + 1.02613S + 0.45485) 0.51352
5 (S2 + 0.27005S + 1.09543) (S2 + 0.70700S + 0.53642) (S + 0.43695) 0.25676
n Numerator of 1/HS) Amax= 0.50dB Denominator (k)

1 S + 2.86278 2.86278
2 S2 + 1.42562S + 1.51620 1.43138
3 (S2 + 0.62646S + 1.14245) (S+0.62646) 0.71570
4 (S2 + 0.35071S + 1.06352) (S2 + 0.84668S + 0.356412) 0.35785
5 (S2 + 0.22393S + 1.03578) (S2 + 0.58625S + 0.47677) (S + 0.362332) 0.17892
n Numerator of 1/H(S) Amax= 1dB Denominator (k)

1 S + 1.96523 1.96523
2 S2 + 1.09773S + 1.10251 0.98261
3 (S2 + 0.49417S + 0.99420) (S+0.49417) 0.49130
4 (S2 + 0.27907S + 0.98650) (S2 + 0.67374S + 0.27940) 0.24565
5 (S2 + 0.17892S + 0.98831) (S2 + 0.46841S + 0.42930) (S + 0.28949) 0.12283
s
Denormalization : S =
p
32
Electronic Filters
Example
600
s = = 3  from graph n = 3
200
From table H(S) normalized is obtained
0.71570
H (S ) =
( S + 0.62646S + 1.14245)(S + 0.62646)
2
Denormalization : S = s
200
Denormalized
5725600
H (s) =
( s 2 + 125.3s + 45698)(s + 125.3)
Bandpass Details
33
Electronic Filters
ELLIPTIC (CAUER) APPOXIMATION
s s
s = H(s) = Loss Function Denormalization : S =
p p
34
Electronic Filters
ELLIPTIC APPOXIMATION
Example
600
s = = 3  from table n = 2
200
From table H(S) normalized is obtained
( S 2 + 17.48528)
H ( S ) = 0.083974 2
( S + 1.35715S + 1.55532)
Denormalization : S = s
200
Denormalized
( s 2 + 699411)
H ( s ) = 0.083974 2
( s + 271.4s + 62212.8)
Bandpass Details
35
Electronic Filters
Phase and delay characteristics
Characteristics of a fourth-order Chebyshev (Amax = 0.5dB);

Characteristics of a fourth-order Butterworth (Amax = 3dB);
(a) Loss, (b) Delay, (c) Step input, (d) Step response.
(a) Loss, (b) Delay, (c) Step response.
36
Electronic Filters
BESSEL APPROXIMATION
H(S) normalized is
Bn (0)
H (S ) =
Bn ( S )
Where Bn(S) is the nth order polynomial which
is defined by the following recursive equation
B0 ( S ) = 1
B1 ( S ) = S + 1
and Bn ( S ) = (2n − 1) Bn−1 ( S ) + S 2 Bn −2 ( S )
Bessel Approximation Function in Normalized and Factored Form
Loss of LP Bessel Approximations
n Denominator Numerator
1 S+1 1
2 S2+3S+3 3
3 (S2+3.67782S+6.45944)(S+2.32219) 15
4 (S2+5.79242S+9.14013) (S2+4.20758S+11.4878) 105
5 (S2+6.70391S+14.2725) (S2+4.64934S+18.15631) (S+3.64674) 945
s
 =  To s = p
 p
Denormalization
p
S = sTo  To =
p
  
2 
2n  1 
2 n −  
p
T  p     p  ( 2 n −1) 
100% = 50  
Delay(%) = e  
To  2n 
 p  Delay of LP Bessel Approximations
37
Electronic Filters
BESSEL APPROXIMATION
Example
Requirements:
a) The delay must be flat within 1 percent of the DC value up to 2KHz.
b) The attenuation at 6KHz must exceed 25dB.
Solution
1. Try a fourth-order Filter. 1% → 0 …  p= 1.9 (from delay
Bessel approximations plots)
 s= (6/2)1.9 = 5.7
Attenuation is only 22 dB   p= 5.7 (from loss Bessel
approximations plots)
2. Try a fifth-order Filter. 1% → 0 …  p= 2.5 (from delay

Bessel approximations plots)
 s= (6/2)2.5 = 7.5
Attenuation is 29.5 dB   p= 5.7 (from loss Bessel
approximations plots)
From table :
945
H (S ) = 2
(S + 6.70391S+ 14.2725) (S + 4.64934S+ 18.15631)(S + 3.64674)
2
Denormalization
p 2.5
To = = = 1.98910−4 sec  S = sTo
 p 2 2 K
3.608108 4.5894108 1.8335104

H ( s) =
s 2 + 3.370 104 s + 3.608108 s 2 + 2.338104 s + 4.5894108 s + 1.8335104 T
Delay(%) = 100% = 0,52%
To f = 2 kHz 38
Electronic Filters
DELAY EQUALIZERS
Function Approximation
VO s 2 − as + b
GAIN = =
VIN s 2 + as + b

=
b
VO N
s 2 − ai s + bi
GAIN = = 2
VIN i =1 s + ai s + bi
The number of delay sections N and their defining

parameters (ai , bi) for approximating a given
delay shape are usually obtained by computer
optimization.
39
Electronic Filters
DELAY EQUALIZERS
Example
Delay Equalization of a fouth-order Chebyshev

(Amax=0.25dB, passband edge = 1 rad/sec)
40
Electronic Filters
FREQUENCY TRANSFORMATIONS
HHP(s)
HP, BP or BR LP HBP(s)
HLP(s)
requirements requirements HBR(s)
Block diagram of the frequency transformation procedure
41
Electronic Filters
HIGH-PASS FILTERS
H HP ( s) = H LP ( S ) S →1000 s3
Requirements H HP ( s) =
s ( s 2 + 103 s + 106 )(s + 103 )
HP LP Normalized
Amin → Amin
Amax → Amax
p → p =1
1K rad/s -> -3dB
p
s → s =
s
✓ H(S)LP Normalized is obtained from LP
500 rad/s -> -18dB
Requirements Normalized
✓ H(S)LP normalized can be transformed
to a high-pass function by the
frequency transformation
p
S=
s
H HP ( s) = H LP ( S ) S → p
s
Example:
HP LP Normalized
Amin = 15dB → Amin = 15dB
Amax = 3dB → Amax = 3dB
 p = 1000 rad / s → p =1
p
 s = 500 rad / s → s = =2
s
Butterworth H ( S ) = 1
LP
( S + S + 1)(S + 1)
2
42
Electronic Filters
BAND-PASS FILTERS
S 2 + 17.48528
Requirements Elliptic H LP ( S ) = 0.083947 2
S + 1.35715 S + 1.55532
BP LP Normalized
Amin → Amin H BP (s) = H LP (S ) S → s 2 +4 2 500000
Amax → Amax 2 500 s
 o = 1 2 =  3 4 p =1 s 4 + 2.12 108 s 2 + 3.89 1014
 4 − 3 H HP ( s) = 0.084
 =  2 − 1 s = s 4 + 4.26 103 s 3 + 5.48 107 s 2 + 8.42 1010 s + 3.89 1014

✓ H(S)BP Normalized is obtained from LP
to a band-pass function by the
s 2 + o
2
S=
s
H BP (s) = H LP (S ) S → s 2 +o 2
s
Example:
Amax = 0.5dB Amin = 20 dB
Passband = 500 Hz to 1000 Hz
Stopbands = DC to 275 Hz and 2000 Hz to 
BP LP Normalized
Amin = 20 dB → Amin = 20 dB
Amax = 0.5dB → Amax = 0.5dB
1 = 2 500 rad / s
→ p =1
 2 = 2 1000 rad / s
 3 = 2 275 rad / s 1818 − 275
→ s = = 3.08
 4 ' = 2 1818 rad / s 1000 − 500
43
Electronic Filters
BAND-REJECT FILTERS
0.98261
Requirements Chebyshev H LP ( S ) =
S + 1.09773 S + 1.10251
2
BR LP Normalized
Amin → Amin H BP ( s) = H LP ( S ) S → 2 5000 s
Amax → Amax s 2 + 4 2 6000000
 o = 1 2 =  3 4 p =1
s 4 + 473741011.3 s 2 + 5.610 1016
 H HP ( s ) = 0.89
 =  2 − 1 s = s 4 + 31279.720 s 3 + 1368935105 s 2 + 7.409 1012 s + 5.610 1016
 4 − 3
✓ H(S)BR Normalized is obtained from LP
to a band-reject function by the
s
S= 2
s + o
2
H BP ( s) = H LP ( S ) S → s
s 2 + o 2
Example:
Amax = 1dB Amin = 20 dB
Passbands = below 1000 Hz and above 6000 Hz
Stopband = 2500 Hz to 3000 Hz
BP LP Normalized
Amin = 20 dB → Amin = 20 dB
Amax = 1dB → Amax = 1dB
1 = 2 1000 rad / s
→ p =1
 2 = 2 6000 rad / s
 3' = 2 2000 rad / s 6000 − 1000
→ s = =5
 4 = 2 3000 rad / s 3000 − 2000
44
Electronic Filters
First-order filters.
45
Electronic Filters
First-order all-pass filter.
46
Electronic Filters
Second-order filtering functions. 47
Electronic Filters
Second-order filtering functions. (continued 1) 48
Electronic Filters
Second-order filtering functions. (continued 2)
49
Electronic Filters
Realization of various second-order filter functions using the LCR resonator: (a) general structure, (b) LP, (c) HP, (d) BP, (e) notch
at 0, (f) general notch, (g) LPN (n  0), (h) LPN as s → , (i) HPN (n < 0).
50
Electronic Filters
REFERENCES
✓ SEDRA, Adel S. and SMITH, Kenneth C. Microelectronic
Circuits. Oxford University Press.
✓ DARYANANI, Gobind. Principles of Active Network
Synthesis and Design. John Wiley & Sons.
✓ LATHI, B. P. Signal Processing & Linear Systems. Berkeley-
Cambridge Press.
✓ RUSTON, Henry and BORDOGNA, Joseph. Electric
Networks: functions, filters, analysis. MacGraw-Hill.
✓ NOCETI FILHO, Sidnei. Filtos Seletores de Sinais. Editora
da UFSC.
51
Electronic Filters

Filtros Eletrônicos

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Filtros Eletrônicos

Uploaded by

Copyright:

Available Formats

ELECTRONIC FILTERS

Celso José Faria de Araújo, Dr.

Classification Classification Classification

Passive Low-Pass Butterworth

Active High-Pass Chebyshev

Digital Band-Pass Eliptic or

c1 o c2  c1 o c2 

• B is the passband width of BP filter

• B is the passband width of BP filter

• B is the passband width of BR filter

• B is the passband width of BR filter

Slope = -40dB/decade Gain

Slope = 40dB/decade Gain

vo (t ) = vin (t − To ) Vo ( s) = Vin ( s) e − sTo

Correction factor :  = 100,1x 0.5 − 1 = 0.35

n Numerator of 1/HS) Amax= 0.50dB Denominator (k)

n Numerator of 1/H(S) Amax= 1dB Denominator (k)

Characteristics of a fourth-order Chebyshev (Amax = 0.5dB);

2. Try a fifth-order Filter. 1% → 0 …  p= 2.5 (from delay

3.608108 4.5894108 1.8335104

The number of delay sections N and their defining

Delay Equalization of a fouth-order Chebyshev

Block diagram of the frequency transformation procedure

You might also like